Author + information
- N. A. Mark Estes III, MD∗ ()
- Tufts University School of Medicine, New England Cardiac Arrhythmia Center, Tufts Medical Center, Boston, Massachusetts
- ↵∗Reprint requests and correspondence:
Dr. N. A. Mark Estes III, Tufts University School of Medicine, New England Cardiac Arrhythmia Center, Tufts Medical Center, 705 Washington Street, Boston, Massachusetts 02111.
With the expansion of indications for implantable cardioverter-defibrillators (ICDs) for primary and secondary prevention of sudden cardiac death, >100,000 implantations of these devices are performed annually in the United States (1,2). Of these, ∼25% are replacements of existing devices because of battery depletion (1,2). The multiple clinical trials defining patient groups benefiting from ICD therapy provide a strong evidence base for initial implantation guidelines developed by professional societies (1). The evidence is considerably less robust related to clinical reassessment to ensure that the patient continues to meet criteria at the time of device replacement (2). Clinical trials comparing outcomes in patients at the time of ICD replacement have not been conducted (2). This lack of empirical data on outcomes after ICD replacement at the time of battery depletion has prevented the development of evidence-based recommendations to guide clinicians (2). In the absence of such guidance, clinicians may feel compelled by ethical or legal considerations to replace the ICD regardless of whether the patient still meets implantation criteria (2). Misaligned economic incentives may influence replacement decisions as well (2). Although the risks associated with ICD replacement have been evaluated, the benefits and costs remain unknown (3). Whether the common practice of replacement without ensuring that guideline criteria are still met is the appropriate clinical strategy remains unknown (2).
Some of the very limited available evidence related to this important issue includes observations in heart failure patients with recovered ejection fractions (4). Outcomes were evaluated in 91 consecutive patients with ICDs placed for primary prevention of sudden death who underwent generator replacement with reassessment of guideline-based implantation criteria. Most of these patients had ischemic heart disease (76%) (4). At generator replacement, 25 patients had an improvement of the left ventricular ejection fraction (LVEF) of at least 10% and to >35% (4). These patients had an LVEF of 49% compared with 32% at baseline (p < 0.0001). During 6.2 ± 2.2 years of follow-up after the original implantation, 9 patients (36%) with an improved LVEF compared with 19 patients (29%) with an unchanged LVEF had appropriate ICD shocks (p = 0.51). The incidence of appropriate ICD shocks was similar between the 2 groups before (p = 0.90) and after (p = 0.97) generator replacement (4). Of the 9 patients with an improved LVEF with appropriate shock, 4 had shocks before generator replacement, 2 had shocks before and after generator replacement, and 3 patients, who never had shocks before, had their first shock after generator replacement (4). On the basis of these observations, the authors conclude that some ICD patients whose LVEF improves to >35% at generator replacement remain at risk of appropriate ICD shocks (4).
Additional data are available in patients with nonischemic cardiomyopathy from the DEFINITE (Defibrillators in Non-Ischemic Cardiomyopathy Treatment Evaluation Trial), which randomized patients with nonischemic cardiomyopathy and an LVEF <36% and frequent ventricular premature contractions or nonsustained ventricular tachycardia to optimal medical therapy alone or with an ICD (5). ICD implantation caused a significant reduction in arrhythmic sudden death (p < 0.006) associated with a trend toward a reduction in total mortality (p < 0.08) (5). An analysis was performed of survival and arrhythmic endpoints in patients whose LVEF was reassessed between 90 and 730 days after enrollment (5). Patients whose LVEF improved had reduced mortality compared with patients whose LVEF decreased (hazard ratio: 0.09, p < 0.001) (5). However, survival free of appropriate shocks was not significantly related to LVEF improvement during follow-up (5). On the basis of these observations, the authors concluded that LVEF improvement was associated with improved survival, but not with a significant decrease in appropriate shocks (5). They also note that these data highlight that appropriate caution should be exercised to not extrapolate the positive effect of improved LVEF to the elimination of arrhythmic events (5).
In this issue of the Journal, additional observations are reported from an investigation by Kini et al. (6) addressing the appropriateness of replacing primary prevention ICDs at time of generator replacement. The investigators performed a retrospective analysis of patients undergoing replacement in the Veterans Affairs hospital system (6). The indications for continued ICD therapy were defined as an LVEF <35% or receiving appropriate device therapy (6). Of the 231 patients analyzed, 59 (26%) no longer met guideline indications for an ICD at the time of generator replacement (6). An additional 79 patients (34%) had not received any appropriate ICD therapies and had not undergone reassessment of their LVEF (6). Patients with an initial LVEF of 30% to 35% were less likely to meet indications for ICD therapy at the time of replacement (odds ratio: 0.52, p = 0.01) (6). Patients without ICD indications subsequently received appropriate ICD therapies at a significantly lower annual rate than patients who with indications (2.8% vs. 10.7%) (6).
There are many limitations of the study including its retrospective design, inclusion of only male veterans, observational rather than interventional nature, and lack of complete clinical data in a minority of patients (6). The cost data were modeled, and a robust cost-effectiveness analysis was not performed (6). They state that their observations make the case to perform ICD explantations instead of generator replacement in patients who experience no appropriate therapies and show significant improvement in the LVEF when their devices reach elective replacement indications (6). In addition, they conclude that in the patients in whom improvement of LVEF has occurred with the original device being cardiac resynchronization therapy (CRT) with a defibrillator, CRT without a defibrillator device could be used instead of CRT with a defibrillator for replacement (6). The authors' conclusions are bold when considered in the context of the limitations of the data and other available evidence. Although this study extends available evidence, the data presented are insufficient to resolve the fundamental issue of the appropriateness of ICD replacement at the time of battery depletion in patients who no longer meet initial indications for ICD implantation.
Other studies examining the time dependence of ICD therapy in patients with primary prevention devices provide appropriate context for consideration of the Kini et al. study (6,7). The incidence and time dependence of first appropriate ICD therapy for ventricular arrhythmia and rapid ventricular tachycardia (cycle length ≤260 ms) have been reported in 525 ICD patients with primary prevention ICDs implanted because of previous MI and LVEF ≤35% (7). Overall, 115 (22%) had appropriate ICD therapy (7). The incidence of first appropriate ICD therapy was highest in the first year post-implantation (20%), decreased to 12% in year 2, and remained at 6% to 11% yearly thereafter for up to 7 years (7). A similar trend was observed with rapid ventricular arrhythmias: a higher risk in the first year (6%) and a lower but persistent risk thereafter (3.8% in year 7) (7). In this patient group with previous MI and an LVEF ≤35%, the incidence of first ICD therapy for ventricular arrhythmias is highest in the first year post-implantation and persists for up to 7 years thereafter (7). The authors concluded that risk of first appropriate ICD therapy persists over time, and thus replacement of ICDs appears to be indicated for all patients (7). However, analysis was not performed on the basis of reassessment of the initial implantation criteria at the time of elective battery replacement (7).
The MADIT II analysis has the robust endpoint of total mortality rather than appropriate ICD therapy use in the Kini et al. (6) and other studies (8). It has been well established that using ICD therapy as a surrogate endpoint results in exaggeration of mortality benefit (9). It has also become evident that although many therapies for ventricular arrhythmias are appropriate, they are unnecessary and increase mortality (10). The MADIT-RIT (Multicenter Automatic Defibrillator Implantation Trial-Reduce Inappropriate Therapy) study demonstrated that programming of ICD therapies for tachyarrhythmias of ≥200 beats/min or with a prolonged delay in therapy at ≥170 beats/min, compared with conventional programming, was associated with reductions in inappropriate and appropriate therapy and all-cause mortality during long-term follow-up (10). Relevant data on time dependence of ICD therapies and mortality in primary prevention with contemporary programming analyzed on the basis of reassessment of ejection fraction after implantation or at the time of replacement are not available (10).
Kini et al. (6) have brought an important clinical issue to the forefront. Their data do support the case to perform ICD explants instead of generator replacement in patients with no appropriate therapies and significant improvement of the LVEF. However, it is evident that there are many gaps in current knowledge related to the optimal ICD replacement strategy. Even with the additional observations of Kini et al., these gaps prevent resolution of this important issue using the standards of evidence-based medicine. Prospective studies are needed of patients at high or low risk of sudden death who are eligible for ICD replacement to identify patient groups that are unlikely to benefit from therapy (2).
In the meantime, it is time for a new approach to elective replacements of ICDs. If the LVEF improves at the time of reevaluation such that a patient no longer meets implantation criteria and has not had appropriate therapy, it is reasonable to have a discussion with the patient regarding the risk of replacement and uncertain benefit. Patients should have a life expectancy of at least 1 year with a reasonable quality of life to justify ICD replacement (1,2). Finally, careful reassessment of the patient's desire for life-sustaining therapies should be undertaken with a informed discussion of their cardiac and general medical status, quality of life, and life expectancy. From a societal perspective, this approach should result in savings (2). Ultimately, in the setting of uncertainty, a deferential approach to patient preference is always a prudent approach (11). Electively withdrawing a potentially life-sustaining therapy such as an ICD by not replacing it is justified and appropriate from both ethical and legal perspectives (11). Although this new approach is reasonable, it is one based on very limited data. Defining the optimal approach will require clinical trials and registries to bridge the many gaps in our current knowledge.
↵∗ Editorials published in the Journal of the American College of Cardiology reflect the view of the authors and do not necessarily represent the views of JACC or the American College of Cardiology.
Dr. Estes is a consultant for Boston Scientific, Medtronic, and St. Jude Medical.
- American College of Cardiology Foundation
- Epstein A.E.,
- DiMarco J.P.,
- Ellenbogen K.A.,
- et al.,
- American College of Cardiology/American Heart Association Task Force on Practice Guidelines (Writing Committee to Revise the ACC/AHA/NASPE 2002 Guideline Update for Implantation of Cardiac Pacemakers and Antiarrhythmia Devices); American Association for Thoracic Surgery; Society of Thoracic Surgeons
- Poole J.E.,
- Gleva M.J.,
- Mela T.,
- et al.,
- REPLACE Registry Investigators
- Schliamser J.E.,
- Kadish A.H.,
- Subacius H.,
- et al.,
- for the DEFINITE Investigators
- Kini V.,
- Khaled Soufi M.,
- Deo R.,
- et al.
- Wilber D.J.,
- Zareba W.,
- Hall W.J.,
- et al.
- Ellenbogen K.A.,
- Levine J.H.,
- Berger R.D.,
- et al.